I am working on an app that utilizes both AR and the ability to speak / get responses. I got halfway on the speech recognition part, though. So I want to know, if someone has experience in this, can you recommend me a good and easy way to setup a voice recognition system?
The idea is to take input through voice and to get a result that you can handle. For example saying “Hi” could generate the response “Hello”. Saying “What is the time?” could utilize the internal clock for a response. I think you get the idea.
Also, I found Google Speech-to-text. But I am not looking for any paid services, since it is just a personal project thing at the moment.
Hi, been there and without online analysis it’s a pain in butt. You can pipe stuff to Google speech API. Interested if anyone has offline solution for this (Windows)
It contains a speech recognition function, but I really need to understand how this works first. There is no documentation and the guy is Japanese I think. If anyone have an idea, I would be grateful for an answer.
Edit: Yeah, about that Google Speech API. Seems like you need to sign up for a pricing solution for that, and for a sole developer who is just doing this to test out, that is probably not what I will go for.
For Android you’ll find plenty of solutions, as those use mainly pocket Sphinx. They love games had earlier word detection in asset store, it’s deprecated now. In that solution you teached the words by recording them, but it was very unreliable.
Long time ago I coded a VB.net version, where I used Google search engine page microphone option for this purpose. That returned the voice as a text, but it required slicing for long sentences.
I haven’t yet found easy way in Unity to do this offline.
I’m also interested in offline solutions, but I’ve just started looking at Watson, which is free for the first 500mins/month. The accuracy of the voice detection is not as good as I thought it would be, but I get that when using siri, and cortana as well, so perhaps it’s my accent.
The CONTUS MirrorFly API structure scales on-demand that allows customers to create voice connectivity prompt with virtually unlimited capacity.
You can build and scale the configurable voice applications in minutes with any form of voice app like the unified communication software, call tracking and call recording software, contact center apps, and whatever that involves voice connectivity.
MirrorFly’s voice API consists of dissimilar APIs that allows you to make activities like translate text with double tap, provide VoIP telecom services, and lastly to add voice and video call functionalities to your website or mobile app. The APIs delivers good documentation that helps you to boost your voice API project instantly.